bio-samtools 0.4.2 → 0.5.0

data/lib/bio/db/{sam/vcf.rb → vcf.rb}

@@ -4,12 +4,14 @@
 # Forked from vcfruby at https://github.com/jesserod/vcfruby 
 # Modified and tests written by Dan MacLean (dan.maclean@tsl.ac.uk)
 # VCF is described at http://www.1000genomes.org/node/101
-
+module Bio
+  class DB
 class Vcf
 public
   attr_accessor :chrom, :pos, :id, :ref, :alt, :qual, :filter, :info, :format, :samples
 
   #create the vcf object, use the ordered list of sample names to label samples if provided ['A', 'B', 'C'], otherwise uses, 1,2,3 etc
+  #vcf = Bio::DB::Vcf("19	111	.	A	C	9.6	.	.	GT:HQ	0|0:10,10	0|0:10,10	0/1:3,3")
   def initialize(line=nil, sample_names=nil)
     @info = {}
     @samples = {}
@@ -20,7 +22,7 @@ public
     Integer.new(x) rescue x
   end
 
-  #gets the info in the Vcf lines and parses the
+  #gets the info in the Vcf lines and parses it, setting the attributes
   def parse_line(line, sample_names=nil)
     return false if line[0,1] == '#'
 
@@ -75,3 +77,5 @@ public
   end
 
 end
+end
+end

data/test/test_basic.rb

@@ -5,7 +5,8 @@ require 'ffi'
 require "test/unit"
 require "bio/db/sam"
 require "bio/db/sam/sam"
-
+require "bio/db/pileup"
+require "bio/db/vcf"
 
 class TestBioDbSam < Test::Unit::TestCase
 
@@ -168,6 +169,7 @@ class TestBioDbSam < Test::Unit::TestCase
     #node_7263       238     60 has 550+, query from 0 to 500, something shall come.... 
   end
 
+=begin ##these tests are correct-ish but the functionality isnt implemented yet ... 
   def test_read_invalid_reference
     sam       = Bio::DB::Sam.new({:bam=>@testBAMFile})
     sam.open
@@ -223,7 +225,7 @@ class TestBioDbSam < Test::Unit::TestCase
       assert(true, "Exception generated and catched")
     end
   end
-
+=end 
   def test_fasta_load_index
     sam = Bio::DB::Sam.new({:fasta=>@testReference})
     sam.load_reference
@@ -289,7 +291,7 @@ class TestBioDbSam < Test::Unit::TestCase
     end
     assert_equal(10,pileup_list.length)
     pileup_list.each  do |p|
-      assert_kind_of(Pileup, p)
+      assert_kind_of(Bio::DB::Pileup, p)
     end
   end
   
@@ -302,13 +304,13 @@ class TestBioDbSam < Test::Unit::TestCase
       list << pile
     end
     assert_equal(10,list.length)
-    list.each {|p| assert_kind_of(Pileup, p)}
+    list.each {|p| assert_kind_of(Bio::DB::Pileup, p)}
     vcf_list = []
     sam.mpileup_plus(:region => "chr_1:100-109", :g => true) do |vcf|
       vcf_list << vcf
     end
     assert_equal(10,vcf_list.length)
-    vcf_list.each {|p| assert_kind_of(Vcf, p)}
+    vcf_list.each {|p| assert_kind_of(Bio::DB::Vcf, p)}
   end
 
   #test whether the call to mpileup returns a vcf object if :g => true is used on the command-line

data/test/test_vcf.rb

@@ -1,17 +1,17 @@
 $: << File.expand_path(File.dirname(__FILE__) + '/../lib')
 $: << File.expand_path('.')
 require 'rubygems'
-require 'bio/db/sam/vcf'
+require 'bio/db/vcf'
 require "test/unit"
 
 
 class TestVcf < Test::Unit::TestCase
   
   def setup
-    @vcf1 = Vcf.new("19	111	.	A	C	9.6	.	.	GT:HQ	0|0:10,10	0|0:10,10	0/1:3,3",["a","b","c"]) #from a 3.3 vcf file
-    @vcf2 = Vcf.new("20	14370	rs6054257	G	A	29	0	NS=3;DP=14;AF=0.5;DB;H2	GT:GQ:DP:HQ	0|0:48:1:51,51	1|0:48:8:51,51	1/1:43:5:-1,-1") #from a 3.3 vcf file
-    @vcf3 = Vcf.new("19	111	.	A	C	9.6	.	.	GT:HQ	0|0:10,10	0|0:10,10	0/1:3,3") #from a 4.0 vcf file
-    @vcf4 = Vcf.new("20	14370	rs6054257	G	A	29	PASS	NS=3;DP=14;AF=0.5;DB;H2	GT:GQ:DP:HQ	0|0:48:1:51,51	1|0:48:8:51,51	1/1:43:5:.,") #from a 4.0 vcf file
+    @vcf1 = Bio::DB::Vcf.new("19	111	.	A	C	9.6	.	.	GT:HQ	0|0:10,10	0|0:10,10	0/1:3,3",["a","b","c"]) #from a 3.3 vcf file
+    @vcf2 = Bio::DB::Vcf.new("20	14370	rs6054257	G	A	29	0	NS=3;DP=14;AF=0.5;DB;H2	GT:GQ:DP:HQ	0|0:48:1:51,51	1|0:48:8:51,51	1/1:43:5:-1,-1") #from a 3.3 vcf file
+    @vcf3 = Bio::DB::Vcf.new("19	111	.	A	C	9.6	.	.	GT:HQ	0|0:10,10	0|0:10,10	0/1:3,3") #from a 4.0 vcf file
+    @vcf4 = Bio::DB::Vcf.new("20	14370	rs6054257	G	A	29	PASS	NS=3;DP=14;AF=0.5;DB;H2	GT:GQ:DP:HQ	0|0:48:1:51,51	1|0:48:8:51,51	1/1:43:5:.,") #from a 4.0 vcf file
   end
   
   

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: bio-samtools
 version: !ruby/object:Gem::Version
-  version: 0.4.2
+  version: 0.5.0
   prerelease: 
 platform: ruby
 authors:
@@ -11,11 +11,11 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2011-11-10 00:00:00.000000000Z
+date: 2012-01-13 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: ffi
-  requirement: &2157872120 !ruby/object:Gem::Requirement
+  requirement: &2153534240 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -23,10 +23,10 @@ dependencies:
         version: '0'
   type: :runtime
   prerelease: false
-  version_requirements: *2157872120
+  version_requirements: *2153534240
 - !ruby/object:Gem::Dependency
   name: bio
-  requirement: &2157870700 !ruby/object:Gem::Requirement
+  requirement: &2153533620 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -34,10 +34,10 @@ dependencies:
         version: 1.4.2
   type: :runtime
   prerelease: false
-  version_requirements: *2157870700
+  version_requirements: *2153533620
 - !ruby/object:Gem::Dependency
   name: shoulda
-  requirement: &2157869680 !ruby/object:Gem::Requirement
+  requirement: &2153531760 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -45,10 +45,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *2157869680
+  version_requirements: *2153531760
 - !ruby/object:Gem::Dependency
   name: bundler
-  requirement: &2157868840 !ruby/object:Gem::Requirement
+  requirement: &2153529440 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -56,10 +56,10 @@ dependencies:
         version: 1.0.0
   type: :development
   prerelease: false
-  version_requirements: *2157868840
+  version_requirements: *2153529440
 - !ruby/object:Gem::Dependency
   name: jeweler
-  requirement: &2157857240 !ruby/object:Gem::Requirement
+  requirement: &2153526840 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -67,10 +67,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *2157857240
+  version_requirements: *2153526840
 - !ruby/object:Gem::Dependency
   name: rcov
-  requirement: &2157856580 !ruby/object:Gem::Requirement
+  requirement: &2153524760 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -78,10 +78,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *2157856580
+  version_requirements: *2153524760
 - !ruby/object:Gem::Dependency
   name: bio
-  requirement: &2157856020 !ruby/object:Gem::Requirement
+  requirement: &2153522840 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -89,10 +89,10 @@ dependencies:
         version: 1.4.2
   type: :development
   prerelease: false
-  version_requirements: *2157856020
+  version_requirements: *2153522840
 - !ruby/object:Gem::Dependency
   name: ffi
-  requirement: &2157855320 !ruby/object:Gem::Requirement
+  requirement: &2153483000 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -100,10 +100,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *2157855320
+  version_requirements: *2153483000
 - !ruby/object:Gem::Dependency
   name: rdoc
-  requirement: &2157854640 !ruby/object:Gem::Requirement
+  requirement: &2153482380 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -111,7 +111,7 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *2157854640
+  version_requirements: *2153482380
 description: ! "Binder of samtools for ruby, on the top of FFI. \n\n  This project
   was born from the need to add support of BAM files to \n  the gee_fu genome browser
   (http://github.com/danmaclean/gee_fu)."
@@ -134,6 +134,7 @@ files:
 - doc/Bio.html
 - doc/Bio/DB.html
 - doc/Bio/DB/Alignment.html
+- doc/Bio/DB/Pileup.html
 - doc/Bio/DB/SAM.html
 - doc/Bio/DB/SAM/Library.html
 - doc/Bio/DB/SAM/Tools.html
@@ -145,6 +146,7 @@ files:
 - doc/Bio/DB/SAM/Tools/SamfileTX.html
 - doc/Bio/DB/SAMException.html
 - doc/Bio/DB/Tag.html
+- doc/Bio/DB/Vcf.html
 - doc/LICENSE_txt.html
 - doc/LibC.html
 - doc/Pileup.html
@@ -241,6 +243,7 @@ files:
 - doc/js/search_index.js
 - doc/js/searcher.js
 - doc/lib/bio-samtools_rb.html
+- doc/lib/bio/db/pileup_rb.html
 - doc/lib/bio/db/sam/bam_rb.html
 - doc/lib/bio/db/sam/faidx_rb.html
 - doc/lib/bio/db/sam/library_rb.html
@@ -248,6 +251,7 @@ files:
 - doc/lib/bio/db/sam/sam_rb.html
 - doc/lib/bio/db/sam/vcf_rb.html
 - doc/lib/bio/db/sam_rb.html
+- doc/lib/bio/db/vcf_rb.html
 - doc/rdoc-style.css
 - doc/rdoc.css
 - doc/table_of_contents.html
@@ -258,20 +262,19 @@ files:
 - ext/mkrf_conf.rb
 - lib/bio-samtools.rb
 - lib/bio/.DS_Store
+- lib/bio/db/pileup.rb
 - lib/bio/db/sam.rb
 - lib/bio/db/sam/bam.rb
 - lib/bio/db/sam/external/COPYING
 - lib/bio/db/sam/external/VERSION
 - lib/bio/db/sam/faidx.rb
 - lib/bio/db/sam/library.rb
-- lib/bio/db/sam/pileup.rb
 - lib/bio/db/sam/sam.rb
-- lib/bio/db/sam/vcf.rb
-- test/basictest.rb
+- lib/bio/db/vcf.rb
 - test/helper.rb
-- test/pileup.rb
 - test/samples/small/ids2.txt
 - test/samples/small/sorted.bam
+- test/samples/small/sorted.bam.bai
 - test/samples/small/test
 - test/samples/small/test.bam
 - test/samples/small/test.fa
@@ -307,7 +310,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: -4357702541244266034
+      hash: 2724404510118255875
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements:
@@ -316,7 +319,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 1.8.6
+rubygems_version: 1.8.10
 signing_key: 
 specification_version: 3
 summary: Binder of samtools for ruby, on the top of FFI.

data/test/basictest.rb

@@ -1,322 +0,0 @@
-$: << File.expand_path(File.dirname(__FILE__) + '/../lib')
-$: << File.expand_path('.')
-require "test/unit"
-require "bio/db/sam"
-require "bio/db/sam/sam"
-
-
-class TestBioDbSam < Test::Unit::TestCase
-
-  #Set up the paths
-  def setup
-    @test_folder                = "test/samples/small"
-    @testTAMFile                = @test_folder + "/test.tam"
-    @testBAMFile                = @test_folder + "/testu.bam"
-    @testReference              = @test_folder + "/test_chr.fasta"
-    
-  end
-  
-  #Removing the index files
-  def teardown
-    begin
-      File.delete(@testReference + ".fai")
-      p "deleted: " + @testReference + ".fai "
-    rescue
-    end
-    begin
-      File.delete(@testBAMFile + ".fai")
-      p "deleted: " + @testBAMFile + ".bai "
-    rescue
-    end
-  end
-=begi
-  def default_test
-    puts $LOAD_PATH
-    assert(true, "Unit test test")
-  end
-
-  def test_openSAMFile
-    bamfile                      = Bio::DB::SAM::Tools.samopen(@testTAMFile,"r",nil)
-    Bio::DB::SAM::Tools.samclose(bamfile)
-    assert(true, "file open and closed")
-  end
-
-  def test_new_class_empty
-    begin
-      bam                        = Bio::DB::Sam.new({})
-      assert(false, "Should fail while opening without parameters")
-    rescue Bio::DB::SAMException => e
-      puts e.message
-      assert(true, e.message)
-    end
-  end
-
-  def test_new_class_empty_invalid_path
-    begin
-      sam                        = Bio::DB::Sam.new({:bam=>"INVALID"})
-      sam.open
-      sam.close
-      assert(false, "Should fail with an invalid path")
-    rescue Bio::DB::SAMException => e
-      puts e.message
-      assert(true, e.message)
-    end
-  end
-
-  def test_class_text_read_no_faidx
-    sam                          = Bio::DB::Sam.new({:tam=>@testTAMFile})
-    sam.open
-    sam.close
-    assert(true, "file open and closed with the class")
-  end
-
-  def test_class_text_read_no_close
-
-    fam                          = Bio::DB::Sam.new({:tam=>@testTAMFile})
-    fam.open
-    fam                          = nil   
-    ObjectSpace.garbage_collect
-
-    assert(true, "file openend but not closed")
-  end
-
-  def test_class_binary_read_no_close
-
-    Bio::DB::Sam.new({:bam=>@testBAMFile}).open
-    ObjectSpace.garbage_collect
-    assert(true, "BINARY file openend but not closed")  
-  end
-
-  def test_read_coverage
-     sam       = Bio::DB::Sam.new({:bam=>@testBAMFile, :fasta=>@testReference})
-     sam.open
-  	File.open( @test_folder +"/ids2.txt", "r") do |file|
-  	  puts "file opened"
-  	  file.each_line{|line|
-        fetching = line.split(' ')[0]
-        puts "fetching: " + fetching
-  	    sam.load_reference
-    	  seq = sam.fetch_reference(fetching, 0, 16000)
-  #	  puts seq 
-  #	  puts seq.length
-  	  als = sam.fetch(fetching, 0, seq.length) 
-     #      p als
-  	  if als.length() > 0 then
-  	       p fetching
-                p als
-  	    end
-           }
-
-  	end
-    sam.close
-   assert(true, "Finish")
-  end
-#  def test_read_TAM_as_BAM
-#    begin
-#      sam                          = Bio::DB::Sam.new({:bam=>@testTAMFile})
-#      sam.open
-#      sam.close
-#      assert(false, "Should raise an exception for reading a BAM as TAM") 
-#    rescue Bio::DB::SAMException => e
-#      assert(true, "Properly handled")
-#    end 
-#  end
-
-# def test_read_BAM_as_TAM
-#    begin
-#      sam                          = Bio::DB::Sam.new({:tam=>@testBAMFile})
-#      sam.open
-#      sam.close
-#      assert(false, "Should raise an exception for reading a BAM as TAM") 
-#    rescue Bio::DB::SAMException => e
-#      assert(true, "Properly handled")
-#    end 
-#  end
-
-  def test_bam_load_index
-    sam       = Bio::DB::Sam.new({:bam=>@testBAMFile})
-    sam.open
-    index = sam.load_index
-    sam.close
-    assert(true, "BAM index loaded")
-    #  attach_function :bam_index_build, [ :string ], :int
-    #  attach_function :bam_index_load, [ :string ], :pointer
-    #  attach_function :bam_index_destroy, [ :pointer ], :void
-  end
-
-  def test_tam_load_index
-    begin
-      sam       = Bio::DB::Sam.new({:tam=>@testTAMFile})
-      sam.open
-      sam.load_index
-      sam.close
-      assert(false, "TAM index loaded")
-    rescue Bio::DB::SAMException => e
-      assert(true, "Unable to load an index for a TAM file")
-    end
-  end
-
-  def test_read_segment
-    sam       = Bio::DB::Sam.new({:bam=>@testBAMFile})
-    sam.open
-    als = sam.fetch("chr_1", 0, 500)
-    p als 
-    sam.close
-    assert(true, "Seems it ran the query")
-    #node_7263       238     60 has 550+, query from 0 to 500, something shall come.... 
-  end
-=end
-  def test_read_invalid_reference
-    sam       = Bio::DB::Sam.new({:bam=>@testBAMFile})
-    sam.open
-    begin
-      als = sam.fetch("Chr1", 0, 500)
-      p als 
-      sam.close
-      assert(false, "Seems it ran the query")
-    rescue Bio::DB::SAMException => e
-      p e
-      assert(true, "Exception generated and catched")
-    end
-  end
-
-  def test_read_invalid_reference_start_coordinate
-    sam       = Bio::DB::Sam.new({:bam=>@testBAMFile})
-    sam.open
-    begin
-      als = sam.fetch("chr", -1, 500)
-      p als 
-      sam.close
-      assert(false, "Seems it ran the query")
-    rescue Bio::DB::SAMException => e
-      p e
-      assert(true, "Exception generated and catched")
-    end
-  end
-
-  def test_read_invalid_reference_end_coordinate
-    sam       = Bio::DB::Sam.new({:bam=>@testBAMFile})
-    sam.open
-    begin
-      als = sam.fetch("chr", 0, 50000)
-      p als 
-      sam.close
-      assert(false, "Seems it ran the query")
-    rescue  Bio::DB::SAMException => e
-      p e
-      assert(true, "Exception generated and catched")
-    end
-  end
-  
-  def test_read_invalid_reference_swaped_coordinates
-    sam       = Bio::DB::Sam.new({:bam=>@testBAMFile})
-    sam.open
-    begin
-      als = sam.fetch("chr", 500, 0)
-      p als 
-      sam.close
-      assert(false, "Seems it ran the query")
-    rescue  Bio::DB::SAMException => e
-      p e
-      assert(true, "Exception generated and catched")
-    end
-  end
-
-  def test_fasta_load_index
-    sam = Bio::DB::Sam.new({:fasta=>@testReference})
-    sam.load_reference
-    seq = sam.fetch_reference("chr_1", 0, 500)
-    p seq 
-    sam.close
-    assert(true, "The reference was loaded")
-  end
-  
-  def test_fasta_load_index
-    sam = Bio::DB::Sam.new({:fasta=>@testReference})
-    sam.load_reference
-    begin
-      seq = sam.fetch_reference("chr1", 0, 500)
-      p "Error seq:"+ seq 
-      sam.close
-      assert(false, "The reference was loaded")
-    rescue Bio::DB::SAMException => e
-      p e
-      assert(true,  "The references was not loaded")
-    end
-  end
-  
-  def test_load_feature
-    
-    fs = Feature.find_by_bam("chr_1", 0, 500,@testBAMFile)
-    
-    p fs
-    assert(true, "Loaded as features")
-  end
-  
-  def test_avg_coverage
-    sam = Bio::DB::Sam.new({:fasta=>@testReference, :bam=>@testBAMFile })
-    sam.open
-    cov = sam.average_coverage("chr_1", 60, 30)
-    p "Coverage: " + cov.to_s
-    sam.close
-    assert(true, "Average coverage ran")
-    assert(3 == cov, "The coverage is 3")
-  end
-  
-
-  def test_chromosome_coverage
-    sam = Bio::DB::Sam.new({:fasta=>@testReference, :bam=>@testBAMFile })
-    sam.open
-    covs = sam.chromosome_coverage("chr_1", 0, 60)
-    p "Coverage: "
-    p covs
-    puts "POS\tCOV"
-    covs.each_with_index{ |cov, i| puts "#{i}\t#{cov}" }
-    sam.close
-    assert(true, "Average coverage ran")
-    #assert(3 == cov, "The coverage is 3")
-  end
-  
-  #test whether the call to mpileup works and returns 10 objects of class pileup
-  def test_pileup
-    sam = Bio::DB::Sam.new(:fasta=>@testReference, :bam=>@testBAMFile )
-    pileup_list = []
-    sam.mpileup(:region => "chr_1:100-1100") do |pile|
-      next unless pile.ref_name == 'chr_1' ##required because in the test environment stdout gets mixed in with the captured stdout in the function and non pileup lines are passed...
-      pileup_list << pile
-    end
-    assert_equal(1000,pileup_list.length)
-    pileup_list.each  do |p|
-      assert_kind_of(Pileup, p)
-    end
-  end
-
-end
-
-class Feature 
-attr_reader :start, :end, :strand, :sequence, :quality
-
-def initialize(a={})
-  p a
-  @start = a[:start]
-  @end = a[:enf]
-  @strand = a[:strand]
-  @sequence = a[:sequence]
-  @quality = a[:quality]
-end
-
-def self.find_by_bam(reference,start,stop,bam_file_path)
-  
-  sam = Bio::DB::Sam.new({:bam=>bam_file_path})
-  features = []
-  sam.open
-  
-  fetchAlignment = Proc.new do |a|
-    a.query_strand ? strand = '+'  : strand = '-'
-    features << Feature.new({:start=>a.pos,:end=>a.calend,:strand=>strand,:sequence=>a.seq,:quality=>a.qual})
-  end
-  sam.fetch_with_function(reference, start, stop, fetchAlignment)
-  sam.close
-  features
-end
-end